Use of Artificial Neural Networks for Prediction of Solar Energy Potential in Southern States of India

Abstract

Prediction and assessment of solar radiation is necessary prerequisite in the setting up and sizing of solar power applications. In this study, an artificial neural network (ANN) model was developed for prediction of solar energy potential in Andhra Pradesh (AP) and Telangana state (TS), India (lies between 12°41' and 22°N latitude and 77° and 84°40'E longitude). Standard multilayered, feed-forward, back-propagation neural networks with different architecture were designed using MATLAB. Geographical and meteorological data of 28 locations in AP & TS for period of recent 22 years from the NASA geo-satellite database were used for the training and testing the network. Geographical parameters (latitude, longitude and altitude), meteorological data (mean sunshine duration, mean temperature, mean wind speed, mean relative humidity and mean precipitation) and the month of the year were used as input data, while the monthly mean solar radiation was used as the output of the network. Statistical error analysis in terms of mean absolute percentage error (MAPE) was conducted for testing data to evaluate the performance of ANN model. The results show that the correlation coefficients between the ANN predictions and actual mean monthly global solar radiation intensities for training and testing datasets were higher than 95%, thus suggesting a high reliability of the model for evaluation of solar radiation in locations where solar radiation data are not available.